Are Datasets optional or needed to use the storage?

[Svenkk4]

Hi,
my question is: Are Datasets optional or needed to use the storage?
I am new to freenas and ZFS and am a bit confused about datasets. It seems that after creating a ZFS volume always a dataset is created for every user, but it is never explained if this is just convenient or necessary to actually use the storage.

 

[danb35]

Datasets aren't necessary. They allow you to divide your pool into multiple portions that act, in many ways, like independent volumes, allowing you to set different permissions, quotas (if any), compression (if any), etc. per dataset, but they aren't necessary.

 

[Svenkk4]

Ok thanks, are datasets share their space with each other or do they need some kind of partitioning (do i need to know how much space each dataset should get or not before i create them)?

 

[no_connection]

Using them might save you from a permission headache later on.
I use one for CIFS as using the main pool directly didn't work for me the first time.

 

[danb35]

Ordinarily datasets share the pool's space. Quotas and reservations can, however, change that behavior. Read up on them before you use them.

 

[BottomsUp]

I am new too and ended up using a data set per CIFS share. I found that if I didn't I could see all data in every share. I believe this is because I used the same permissions for each share but my use case is that I want mount points on my windows box for media, photos, video. This inhale 3 data sets with those and then share them each.

Is this best practice? I find that data sets are like mount points based on what I am used to in Linux.

 

[danb35]

I think the analogy to mount points is a good one, and it's one that's used pretty regularly when ZFS gurus are talking about it. The biggest difference that I see is that mount points under Linux are generally partitions of fixed size, while datasets don't have to be (though you can give them quotas to effectively fix their sizes as well).